produse study_full text

8/10/2019 PRODUSE Study_Full Text

1/156

Productive Use of Energy PRODUSEMeasuring Impacts of Electrification on Small and Micro-Enterprises

in Sub-Saharan Africa


2/156

Copyright 2013

Deutsche Gesellschaft fr Internationale Zusammenarbeit (GIZ) GmbH

P.O. Box 5180

65726 Eschborn, Germany

[email protected]

All rights reserved

The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors.

Photos

GIZ: Moussa Doumbia, Caroline Heidtmann, Marco Hls, Kamikazz, Lucius Mayer-Tasch,

Christoph Messinger, Monika Rammelt

Shutterstock

Layout

creative republic

Thomas Maxeiner Visual Communications

Frankfurt, Germanywww.creativerepublic.net

Language Editor

Sina Mabwa

Acknowledgments

The present study was implemented by a joint Task Team under the supervision of Lucius Mayer-Tasch (GIZ),

Mohua Mukherjee (World Bank) and Kilian Reiche (lead consultant) with funding from the German Federal Min-

istry for Economic Cooperation and Development (BMZ) and the Energy Sector Management Assistance Program

(ESMAP). The studys main authors are listed at the start of each chapter. In addition, we would like to thank the

following contributors, reviewers and interviewed experts: Benjamin Attigah, Sudeshna Ghosh Banerjee, Anna

Brderle, Elizabeth Elizondo, Mike Enskat, Raluca Golumbeanu, Maya Hirsch, Marco Hls, Ulrich Laumanns,

Suzanne Maia, Irene Nakiwu, Alain Ouedraogo, Dana Rysankova and William Steel.

Productive Use of Energy (PRODUSE) is a joint initiative of the Energy Sector Management Assistance Program

(ESMAP), the Africa Electrification Initiative (AEI), the EUEI Partnership Dialogue Facility (EUEI PDF) and Deutsche

Gesellschaft fr Internationale Zusammenarbeit (GIZ). Further information onwww.produse.org.

The financial and technical support by the Energy Sector Management Assistance Program (ESMAP) is grate-

fully acknowledged. ESMAP is a global knowledge and technical assistance program administered by the World

Bank that assists low- and middle-income countries to increase their know-how and institutional capacity to

achieve environmentally sustainable energy solutions for poverty reduction and economic growth. ESMAP is

funded by Australia, Austria, Denmark, Finland, France, Germany, Iceland, Lithuania, the Netherlands, Norway,

Sweden, the United Kingdom, and the World Bank Group.


3/156

Edited by Lucius Mayer-Tasch, Mohua Mukherjee and Kilian Reiche

Productive Use of Energy PRODUSEMeasuring Impacts of Electrification on Small and Micro-Enterprisesin Sub-Saharan Africa


4/156

4

Table of Contents

List of Tables and Figures ....................................................6

List of Acronyms .....................................................................8

Executive Summary .............................. ............................... 10

Chapter 1:Introduction......................................... 15

Chapter 2:The Impact of Electricity Accesson Economic Development:A Literature Review............................ 19

1. Introduction .......................................................................20

2. Conceptual Background: The Steps from

Electricity Supply to Poverty Reduction .................. 21

3. Methodological Issues ............................... .................... 21

4. Macro-Level Research................................. ....................23

4.1. Introduction ............................... .......................... 23

4.2. Empirical Evidence .............................. ...............23

5. Micro-Level Research ................................. .................... 28

5.1. Introduction .............................. .......................... 28

5.2. Factors Influencing Impacts of Energy .... 28

5.3. Empirical Evidence ............................. ............... 29

6. Summary and Conclusion of the

Literature Review ................................. ........................... 36

Chapter 3:Methodology........................................ 39

1. Potential Linkages between Electricity

and Income Generation ................................................40

2. The Treatment: Availability and

Connection .........................................................................40

3. Identification Strategies ............................ ................... 42

3.1. The Identification Problem ........................... 42

3.2. Cross-Sectional Impact Evaluation .............43

3.3. Evaluation Strategies after an Ex-Post

Survey ................................ ................................. .... 44

4. Study Implementation ............................. .....................47 4.1. Selection of Appropriate Control

Regions ...................................................................47

4.2 Survey Approach ................................. ............... 48

Chapter 4:Electrification and Firm Performancein Rural Benin:An Ex-Ante Impact Assessment...... 51

Abstract ...............................................................................52

1. Introduction ........................................................................52

2. Data Collection..................................................................53

3. Research Approach and Focus ....................................54

3.1. Identification Strategy .............................. .......54

3.2. Electricity Dependency: Reliant and

Non-Reliant Enterprises ..................................55

4. Economic Conditions in the Survey

Regions ................................. ................................. ............ 56

4.1. Business Environment and

Infrastructure.................................... .................. 56

4.2. Micro-Finance and BDS ................................ .. 56

4.3. Market Access ............................. .........................58

5. Intermediate Outcomes: Electrification Impacts

on Inputs .............................................................................60

5.1. Energy Usage .............................. ........................60

5.2. Employment ............................... ......................... 62

5.3. Capital ................................ ................................. ... 63

6. Impacts on Firm Performance ................................ ...64

6.1. Profits in the Service Sector .......................... 64

6.2. Profits in the Manufacturing Sector ......... 65

6.2.1. Difference in Means .............................. 65

6.2.2. Matching Approaches..........................68

7. Microeconomic Considerations of

Firm Behavior .............................. ................................. ..... 708. Conclusions ........................................................................72

Chapter 5:Firm Performance and ElectricityUsage in Small Manufacturingand Service Firms in Ghana............. 75

Abstract ............................ ................................. ................. 76

1. Introduction ................................ ................................. ...... 76

2. Empirical Strategy and Data .......................................77

2.1. Sampling and Survey Design .......................... 77 2.2. Potential Impacts and Outcome

Indicators ................................. ..............................78

2.3. Identification Strategy ............................. ........78


5/156

5

3. Economic Conditions in Surveyed Area.................. 79

3.1. Access to Infrastructure ............................ ..... 79

3.2. The Surveyed Firms: Access to Markets,

Finance and Business Development

Services ............................ ................................. .....804. Intermediate Impacts of Electricity Usage .......... 82

4.1. Energy Sources .............................. ..................... 82

4.2. Lighting ................................ ................................. .83

4.3. Energy Expenditures ............................. ............85

4.4 . Capital and Employment .............................. ..85

5. Impact Analysis.................................................................87

5.1. Sales and Firm Revenue ................................. . 88

5.2. Firm Profits: Difference in Means ............... 88

5.3. Ordinary Least Squares ................................. ..89

5.4 . Two Stage Least Squares .............................. ...91

5.5. Robustness of Results ............................. .........91

6. Conclusion ................................ ................................. ........ 92

Chapter 6:Micro-Enterprise Electricity Usagein Two Export-Oriented FishingCommunities at Lake Victoria,Uganda.................................................... 95

Abstract ..............................................................................96

1. Introduction .......................................................................96

2. Empirical Strategy and Data ......................................96

2.1. Sampling and Survey Design ....................... 96

2.2. Treatment, Comparability Issues andIndicators ............................. ................................. 97

3. Economic Conditions in the Surveyed Area ......... 98

3.1. Roads and Energy Infrastructure ................ 98

3.2. Access to Finance and Business

Development Services ............................... ......99

3.3. The Typical Enterprise in the Survey

Area .........................................................................99

4. Quantitative Evidence: Energy and

Enterprise Outcomes ....................................................99

4.1. Energy Usage ............................... .......................99

4.2. Capital ............................ ................................. ......103

4.3. Employment and Operation Hours ......... 104 4.4. Market Access ............................. .......................105

4.5. Firm Profits................................................... ...... 106

4.6. Industry Structure and Product

Variety ................................ ................................. ..107

5. Qualitative Evidence: Stylised Case Studies ....... 108

5.1. Electricity Not Required: Carpentry and

Upholstery............................................... ........... 108

5.2. Electricity is Crucial: Aquaculture ............. 109

5.3. Electricity Improves Offered Services:Barber, Phone and Computer Shop.......... 109

6. Conclusion ........................................................................110

Chapter 7:Conclusions......................................... 113

Annex:Impact Monitoring and Evaluationof Productive Electricity Use

An Implementation Guide forProject Managers............................... 119

1. Introduction .............................. ................................. ...... 120

2. Classical M&E vs. Impact M&E ................................. 121

2.1. Outcomes, Impacts and Highly-

Aggregated Impacts ........................................121

2.2. Second Round Effects .............................. ....... 122

3. Developing a Productive Use Impact M&E

System .............................................................................122

3.1. General strategies to isolate the projects

effect .....................................................................122

3.2. Three PUE Impact M&E Modules .............. 123

4. Step-by-step Towards an Effective PUE ImpactM&E System ....................................................................126

5. Aid Items ............................................................................136

Aid 1. The Results Chain Concept and

Demarcation Between Outcomes

and Impacts ...................................................136

Aid 2. Strategies to Identify the

Counterfactual Situation ......................... 136

Aid 3. List of Indicators ............................ ...............139

Aid 4. Outline of Terms of Reference for

Short-Term Experts ........... .......................... 142

Aid 5. Outline of Inception Report .................... 143

Addenda............................................... 145

References........................................... 146


6/156

6

List of Tables and Figures

Table 1: Effects of Infrastructure and Energy on Productivity ............................... ................................. ......................... 24

Table 2: Effects of Infrastructure and Electricity on Economic Growth ............................. ................................. .........26

Table 3: Effects of Infrastructure on Poverty Reduction in Developing Countries ............................... .................... 27

Table 4: Effects of Electricity Access on Number of Businesses ................................. ................................. ....................29

Table 5: Effects of Electricity Access/Quality on Firm Productivity ................................ ................................. ...............30Table 6: Effects of Electricity Access on Employment .............................. ................................. ................................. .......... 32

Table 7: Effects of Electricity Use on Income at Firm and Household Level ............................... ................................ 34

Table 8: Effects of Electricity Use on Poverty ............................... ................................ ................................. ........................... 35

Table 9: Advantages and Disadvantages of Different Impact Evaluation Approaches ............................... ......... 46

Table 10: Sampling ................................. ................................ ................................. ................................. ................................. ........... 53

Table 11: Electricity Reliant and Non-Reliant Firms in the Access Region .............................. ................................. ...... 55

Table 12: Subjective Reasons for Not Applying for Credits ................................ ................................. ................................ . 57

Table 13: Association Membership, Cooperation and BDS ............................... ................................. ................................. . 57

Table 14: Entrepreneurs Perception of BDS ............................ ................................. ................................. ................................. .58

Table 15: Major Problems as Mentioned by the Entrepreneurs (multiple answers possible) ............................. .58

Table 16: Location of Product Selling Manufacturing Firms (multiple answers possible) ................................ ..59

Table 17: Destination of Products Manufacturing Firms .............................. ................................. ................................. ..59

Table 18: Artificial Lighting Usage Among Electricity Non-Reliant Firms ............................ ................................. .......60

Table 19: Energy Using Appliance Usage (electric and other) ............................. ................................. .............................. 61

Table 20: Connection Rates and Electricity Consumption .............................. ................................. ................................. ... 61

Table 21: Employment and Wages ................................ ................................. ................................. ................................ ...............62

Table 22: Average Capital Stock per Firm in the Service Sector (in FCFA) ............................ ................................. .........63

Table 23: Capital Usage in the Manufacturing Sector..................................... ................................. ................................. ... 64

Table 24: Profits in the Service Sector, Access and Non-Access Region ................................ ................................. .........65

Table 25: Profits in the Service Sector, Connected and Non-Connected Firms............................................................65

Table 26: Mean Performance Indicators by Access and Non-Access Region ................................. ............................... 65

Table 27: Mean Performance Indicators in the Access Region ................................ ................................. ......................... 66

Table 28: Electricity Reliant and Non-Reliant Manufacturing Firms in the Access Region ................................... 66

Table 29: Contribution of Electricity-Reliant Firms to the Local Economy.......................................... ........................... 67

Table 30: Mean Performance Indicators by Region, Excluding Electricity-Reliant Firms ............................... ..........67Table 31: Mean Profits in the Access Region, Excluding Electricity-Reliant Firms ................................ ......................67

Table 32: Probit Regression Grid Connection as Dependent Variable .............................. ................................ .......... 69

Table 33: Mean Profits for Connected and Hypothetically Connected Firms .............................. ............................... 69

Table 34: Subcategories in the Service and Manufacturing Sector ................................ ................................. ................ 77

Table 35: Integration with External Markets (only manufacturing firms) .............................. ................................. .... 80

Table 36: Access to Finance ................................. ................................. ................................ ................................. .............................81

Table 37: Subjective Reasons For Not Applying For Credits ............................... ................................ ................................. ..81

Table 38: BDS Usage .............................. ................................. ................................ ................................. ................................. ............82

Table 39: Connection Rates in Selected Business Types ............................... ................................. ................................. .......82

Table 40: Reason for Not Connecting to the Grid (multiple answers possible) ............................. ............................. 83

Table 41: Lighting ............................ ................................. ................................. ................................. ................................. ..................84

Table 42: Cost Reduction Through Electric Lighting ................................ ................................. ................................. .............84Table 43: Monetary Expenditures and Time Spent to Obtain Respective Energy Source ............................... ........85

Table 44: Resale Value of Capital Endowment (in GH) .............................. ................................. ................................. ....... 86

Table 45: Daily Operation Hours ............................... ................................. ................................ ................................. ................... 86


7/156

7

Table 46: Labor Input per Week .............................. ................................. ................................. ................................. ...................... 87

Table 47: Characteristics of Workforce and Compensation.................................... ................................. ............................ 87

Table 48: Revenue per Month (in GH) ................................ ................................. ................................. ................................ .......88

Table 49: Labour Productivity: Revenue (in GH) per Total Input of Working Hours ................................................ .88

Table 50: Profits (in GH) Difference in Means ............................. ................................. ................................. ...................... 89Table 51: Monthly Firm Profit Ordinary Least Squares and Two Stage Least Squares Regression

(p-values in parentheses) ..............................................................................................................................................90

Table 52: Monthly Profits Manufacturing and Service Sector (including outliers) .............................. ..................92

Table 53: Average Firm Characteristics in the Survey Area .............................. ................................. ................................ 100

Table 54: Reasons for Not Connecting .............................. ................................. ................................. ................................. ..... 100

Table 55: Share of Firms Using the Respective Energy Source (in %) ................................ ................................. .............101

Table 56: Energy Expenditures in UShs ............................. ................................. ................................ ................................. .......102

Table 57: Lighting Sources ................................ ................................ ................................. ................................. ............................. 102

Table 58: Capital Endowment .............................. ................................ ................................. ................................. ........................ 103

Table 59: Hours of Operation and Monthly Labour Input .............................. ................................. ................................. ..104

Table 60: Employment and Remuneration ............................... ................................. ................................. .............................. 105

Table 61: To Whom of the Following Does the Enterprise Sell the Majority of its Goods and Services to?

(in %) ............................... ................................. ................................. ................................. ................................. .................. 106

Table 62: In Which of the Following Places Are the Majority of this Enterprise's Goods and Services

consumed? (in %) ................................. ................................. ................................. ................................. ........................ 106

Table 63: Calculated Profits in an Average Month ............................... ................................. ................................. .............. 106

Table 64: Industry Structure (in %) ............................. ................................. ................................. ................................. ............... 107

Table 65: Products Sold (in %) ................................. ................................. ................................. ................................. .....................108

Table 66: Potential Approaches for PUE impact M&E An Overview ............................... ................................. ........... 124

Figure 1: The Theory: Steps from Electricity Supply to Poverty Reduction ................................. ................................. .. 22

Figure 2: Macro-Level Correlation Between Electricity and Human Development ................................................ .. 22Figure 3: Contribution of Infrastructure to Total Factor Productivity of Firms .............................. .............................. 31

Figure 4: Pathways from Electricity to Income Generation..................................................................................................41

Figure 5: The Difference-in-Differences Approach ............................ ................................. ................................. ................... 45

Figure 6: Access and Non-Access Region .............................. ................................. ................................. ................................. ...54

Figure 7: Matching of Hypothetically Connected Firms ............................ ................................. ................................. ........ 68

Figure 8: Optimal Production Level by Technology ................................. ................................ ................................. ................ 71

Figure 9: Steps in the Design of the PUE Impact M&E Approach ............................... ................................ ....................127

Figure 10: Exemplary Results Chain of Grid Extension Programme ............................ ................................. .................... 137


8/156

8

List of Acronyms

2SLS Two Stage Least Squares

ADB Asian Development Bank

AFD Agence Franaise de Dveloppement

AMFIU Association of Microfinance Institutions of Uganda

BDS Business Development ServiceBMZ German Federal Ministry for Economic Cooperation and Development

BUDS Business Uganda Development Scheme

CIA Conditional Independence Assumption

DD Difference-in-Difference-Estimation

DFID Department for International Development

DGIS Directorate-General for International Cooperation of the Dutch Ministry of Foreign Affairs

EIU Economist Intelligence Unit

EnDev Energising Development programme

EnPoGen Energy, Poverty and Gender

ESMAP Energy Sector Management Assistance Program

EU European Union

EUEI PDF EUEI Partnership Dialogue Facility

FCFA West Africa Franc

GDP Gross Domestic Product

GH Ghanaian New Cedi

GIZ Deutsche Gesellschaft fr Internationale Zusammenarbeit (GIZ) GmbH

GTZ Gesellschaft fr Technische Zusammenarbeit (GTZ) GmbH

HBS Household Budget Survey

HDI Human Development Index

ICE ICE Energy

ICT Information and Communications Technology

IEA International Energy Agency

IEG International Energy Group

ISSER Institute of Statistical Social and Economic Research

IT Information TechnologyIV Instrumental Variables

LSMS Living Standards Measurement Study

M&E Monitoring & Evaluation

M&EED Group Monitoring and Evaluation in Energy for Development

MDG Millennium Development Goal

MENA Middle East North Africa

MPI Multidimensional Poverty Indicator

MSME Micro, Small and Medium Enterprises

NED Northern Electricity Department

NGO Non-Government Organisation

NONIE Network of Networks on Impact Evaluation

NRECA National Rural Electric Cooperative AssociationOLS Ordinary Least Squares

PRODUSE Productive Use of Energy

PUE Productive Use of Electricity

R&D Research & Development


9/156

9

SBEE Small Business Energy Efficiency

SE4All Sustainable Energy for All initiative

SHS Solar Home Systems

SME Small and Medium Enterprise

SPSS Predictive analysis softwareSTATA Data analysis and statistical software

TC Total Cost

TFP Total Factor Productivity

ToR Terms of Reference

TR Total Revenue

U.S. United States of America

UK United Kingdom

UMEME Electricity distribution company in Uganda

UN United Nations

UNDP United Nations Development Programme

US $ United States Dollar

UShs Uganda Shillings

VRA Volta River Authority

WHO World Health Organization


10/156

10

Executive Summary

While the interest of policy makers in the nexus between electrification, productive electricity usage anddevelopment impacts has been increasing steadily over the last decade, the lack of robust evidence on causal

effects of electrification is striking. The joint GIZ-ESMAP study Productive Use of Energy (PRODUSE) Measuring

Impacts of Electrification on Small and Micro-Enterprises in Sub-Saharan Africaset out to improve the under-

standing of this issue. PRODUSE pursued two main objectives: (a) gaining insights on the interaction between

electrification and productive electricity usage by examining the impact of electrification on micro-enterprises

and (b) improving the available toolkit for the impact evaluation of electrification programmes.1

PRODUSE has shown that proper usage of statistical techniques is required for deriving solid findings on

these impacts and has demonstrated that methodological rigour is possible even if available project evalua-

tion budgets are small. The study has confirmed that the ex-ante differences between firms that get

connected to electricity and those that do not get connected are substantial which invalidates any determi-

nation of impacts by simply comparing these groups using descriptive statistics (as is all too often done in

literature on electrification impacts). Methods have to be used which account for observable and also for non-

observable heterogeneity between connected and non-connected firms.

With regards to objective (a) i.e. gaining insight on the interaction between electricity access, productive electric-

ity usage, income generation and additional services, valuable and partly surprising findings could be provided

based on field surveys in Benin, Ghana and Uganda, in spite of the modest budget of the surveys. Stark differ-

ences between industries show up: while service firms tend to get connected to the grid, take-up rates in the

manufacturing sector of rural areas were low in the countries that have been studied. Connected firms in rural

areas in both the manufacturing and the service sectors use electricity mostly for lighting and phone charging.

Some rural manufacturing firms also use electric appliances if it is essential for their production process (such

as welders). In general, however, take-up of electric appliances remains modest. In the service sector more appli-

ances are used, mostly refrigerators and entertainment devices. A slightly different picture prevails in the peri-

urban set-up studied in Ghana. Here, grid connected firms employ much more electric machinery.

Altogether, in the three studies electricity usage did hardly translate into higher firm profits in a measurable

way.2 In one country case, Benin, it seems that the financial burden resulting from the investment in the

connection and subsequent electricity bills can even reduce the profitability of firms, indicating that from a

pure business perspective getting connected is not always a rational option.

These rather sobering results (i.e. generally no clear indication for positive effects of electricity access on firm

performance) were contrasted by some evidence indicating that electrification can lead to the creation of

new firms, which generate additional income and, hence, impacts on the target population in the project

regions. Small service and manufacturing firms are created offering goods and services that have previously

been imported from other regions or simply not been offered in the area heretofore. In addition, individual

cases could be observed, in which larger firms were attracted to the region by the availability of electricity.While such direct investments could contribute substantially to income generation in the region, it is prema-

ture to claim that such firm creation occurs systematically. More research in other regions and with larger

sample sizes is needed to further understand this potential process of electricity-induced firm creation and

1) Regarding objective (b), our aim was to demonstrate evaluation methods which would (i) allow for more robust results than most electrifi-

cation evaluations to date, and at the same time (ii) be readily applicable in real-life implementation contexts which often face limitations

in terms of costs and/or timing. The impact evaluation methods we propose could be applied with relatively modest additional effort to

most electrification programmes. As an example, we provide a next best method to treat project implementation cases in which no

baseline has been established by the time an evaluation starts (which should obviously be avoided wherever possible, but all too often

happens in practice).

2) One can think of a whole series of possible explanations for this result of our three case studies, such as lack of access to external markets,

lack of business skills, etc. One possible explanation that has repeatedly been brought forth by this studys peer reviewers is the low

reliability of the electricity grid. However, none of our three case studies allows for clear conclusions regarding these explanations. For

example, the grid in the surveyed region in Northern Benin, was stable with short outages occurring only once every few days. In Ghana

and Uganda, both announced and unannounced outages occurred somewhat more frequently, but even here only a small number of entre-

preneurs complained about reliability issues. Also, only few non-connected firms declared reliability as a major reason for not connecting.


11/156

11

3) See Chapter 3 (methodology) and the PRODUSE Impact M&E Guide in the annex.

4) The PRODUSE website(www.produse.org) is available as a platform for making available studies that fulfill these requirements.

investments. This particularly includes studies that survey project target regions before and after electrifica-tion and compare firm creation at these two points in time, respectively. Furthermore, crowding-out effects

(i.e. people have to reduce their expenditures for the oldproduct in order to buy the newone) have to be taken

into account in order to assess the net benefit for a region.

Methodologically, PRODUSE has developed and applied a solid approach for gaining insight on how micro-

enterprises use electricity and the extent to which this changes their production process. In spite of this

innovative contribution, PRODUSE cannot be more than a kick-off to further and broader investigations of the

complex relationship between electricity access and productive processes and, eventually, economic develop-

ment and poverty alleviation. It can be concluded that cross-sectional methods if properly implemented

are a valid approach to identify causal effects of electrification on micro-enterprises. Furthermore, the ex-ante

cross-sectional approach generates insights into firm characteristics and behaviour in a comparable, already

electrified non-project region as well as in the project region that can inform the design and implementation

of the planned electrification project. For example, the baseline data from the already electrified control group

can be used during project implementation for developing realistic business plans together with firms in the

project area.

Nonetheless, it would be desirable to also collect over-time data in order to allow the application of differ-

ence-in-differences analysis. In contrast to cross-sectional data, this accounts for unobservable heterogeneity

between connected and non-connected firms, which in turn increases the robustness of results. Furthermore,

it would be desirable to have bigger sample sizes in future studies, because the heterogeneity of firms and

their responses is so high that small sample sizes are often not able to grasp potentially existing differences

in a statistically significant way (even if the survey is focused on specific industries). In addition, the scope of

research might be extended to all sectors in one region and also neighbouring communities in order to

capture demand movements and, hence, crowding out effects.

We strongly encourage development practitioners and policy makers to make use of rigorous evaluation

methodologies such as the one used for PRODUSE3when planning new energy interventions to i) improve

project results and ii) contribute to a more solid overall understanding of the nexus between electrification,

productive use and development impacts. As the literature review(Chapter 2) has shown, there are very few

solid studies on this topic to date. Once a critical mass of robust evaluation studies has been conducted in a

sufficiently broad variety of country, market and project contexts, it will be possible to draw more general

conclusions about this nexus.4

One of the conclusions that can be drawn from the research efforts presented in this report is that project

managers should be realistic in their expectations with regard to the (measurable) economic impact of elec-

trification projects, especially on firms. If substantial productive take-up is specifically intended by an electri-

fication project, a typical strategy would be to include the major determinants for productive uptake in theprogrammes geographic area targeting process (i.e. picking those areas first that appear to be most promis-

ing for productive uses for example because of better access to external markets). However, this may be in

direct contrast to other selection criteria (such as poverty targeting).


12/156

12

The PRODUSE Manual, which has been developed in parallel to this study, provides guidance on how to design

and implement activities promoting productive use that can be integrated into broader electrification

projects and enhance the impact of electrification projects and programmes on local economic development

in general and firm productivity in specific. However, the results of our study show that productive use is notautomatically associated with positive impacts on firm performance and other parameters. Promotion

activities should therefore include support for proper business plan development5for the targeted firms (i.e.

the potential commercial electricity customers) in order to ensure the profitability of their investment into

grid connections and electric appliances. Such promotion activities have to be open towards the results: Con-

necting to the grid should not be promoted at all costs. The decision should rather be based on the business

plan implying that the recommendation for an individual firm can as well be to abstain from a connection if

the projected additional revenue is insufficient to recover the investment. This is essential in order to avoid

predictable misallocations, which might drive some firms into financial problems, as appeared to have hap-

pened in the case study from Benin (electrification trap). Furthermore, the creation of promising new

enterprises as observed in Benin and Uganda could be facilitated by accompanying activities that support

potential external investors in collecting the required information to prepare firm creation in the region.

This could be done in cooperation with industry chambers or regional development programmes.

On a more general note, the findings of the PRODUSE study suggest that (rural) electrification should

not be reduced to its potential contribution to productive uses and, hence, to economic growth in a

narrower sense. Firstly, this poses the risk that claimed objectives are not achieved, as productive uptake

can be moderate in the short term, as our country cases show. Secondly and more importantly, this

would neglect the non-productive significance that electricity arguably has to people in rural areas.

From the perspective of rural dwellers, electric lighting, television and mobile phone charging revolu-

tionise their lives. In this context, it should not be forgotten that productive use in specific and

economic growth in general are only proxies to measure improvements in peoples well-being. Elec-

tricity and modern energy services at large, however, directly affect the well-being of rural people

beyond any potential income generation. In the same vein, the UN has recently included electric-

ity access explicitly as a direct indicator of their new Multidimensional Poverty Indicator (MPI).6

5) See modules 5.3and5.4of the PRODUSE manual Productive Use of Energy (PRODUSE)

A Manual for Electrification Practitioners,which has been developed by GIZ and EUEI PDF.

It can be accessed at www.produse.org/manual.

6) The MPI is based on the Human Development Index (HDI) and formulates ten dimensions

that capture poverty.


13/156

13


14/156


15/156

Chapter 1:Introduction

By Kilian Reiche and Jrg Peters


16/156

16

Introduction

Access to modern energyis considered key to poverty alleviation and growth strategies by policy makers anddevelopment practitioners alike. While energy access was not declared an explicit Millennium Development

Goal (MDG) in 2000, UN (2005) emphasized the role that modern energy plays in achieving most of the

MDGs; UN (2010) stated that energy is at the heart of most critical economic, environmental and develop-

mental issues facing the world today; and the recent SE4All Initiative has mobilized broad international

support for achieving universal energy access with good chances to make energy access an explicit UN Devel-

opment Target in the Post-2015 Development Agenda (UN 2013, UN SDSN 2013).

Access to electricity in particular ranks high in subjective demand prioritizations in most household surveys.

According to the SE4All Global Tracking Framework (2013), about one billion people need to be provided with

electricity to achieve universal access by 2030 at total investments of about US $ one trillion.

At the same time, the energy sectors in virtually all least developed countries are in need of urgent improve-

ments: generation capacities are often not sufficient (with frequent blackouts as a consequence) and electric-

ity tariffs are hardly cost covering, making the extension of electricity grids difficult. New electricity users in

rural areas are the most unattractive market segment, due to low demand densities and a relatively higher

fraction of low income households compared to connected areas. As a response to these difficult conditions,

governments as well as bilateral and multilateral donors have increased their efforts in this sub-sector over

the last decade and subsidized grid and offgrid electrification in many countries.

Electrification practitioners often refer to a set of common sense-based causalities to underpin the hypothe-

sis that electrification contributes to poverty alleviation. Most notably, they often emphasize the crucial role

productive electricity take-upmay play in increasing income generation for home businesses and enterprises

- thus making electricity service provision more viable in turn. Accordingly, UN (2010) and SE4All (2013) both

stress the importance of productive energy uses as an explicit element of universal access strategies.7

However, systematic evidence on (i) the relationship between access to electricity and poverty alleviation in gen-

eral as well as (ii) productivity impacts of electrification in specific are scarce, and rigorous evaluations of electri-

fication interventions particularly in Africa are virtually non-available(IEG 2008, Estache 2010, Chapter 2).8

In order to (i) start filling this gap between practitioners perceptions and the lack of robust evidence and

to (ii) inform the design of interventions to promote the productive use of energy, GIZ and ESMAP launched

the Productive Use of Energy (PRODUSE) study. In three African countries Benin, Ghana and Uganda the

usage of electricity in micro-enterprises was examined and the effect of electricity use on firm perfor-

mance was assessed.

At the outset of the project, a joint survey methodology for all cases was developed: a number of firms should

be surveyed that would be sufficiently high to allow the application of statistical tools. In the optimal case,both an electrified and a non-electrified area are included in order to obtain an idea about how the perfor-

mance of firms differs taking into account other differences like, for instance, the industry they belong to,

the firm size or the entrepreneurs education. The methodology aims to identify counterfactual outcomes

and, thereby, the PRODUSE study is one of few attempts to date to apply rigorous evaluation techniques (as

outlined in Ravallion 2008) to analyse the particular impact of electrification on micro-enterprises.

7) UN 2010: The global community should aim to provide access for the 2-3 billion people excluded from modern energy services, to a basic

minimum threshold of modern energy services for both consumption and productive uses.SE4All (2013): The SE4ALL universal access goal

will be achieved only if every person on the planet has access to modern energy services provided through electricity, clean cooking fuels, clean

heating fuels, and energy for productive use and community services.

8) IEG 2008: The evidence base remains weak for many of the claimed benefits of [rural electrification]. Tailor-made surveys, designed to test

these benefits, need to be built into a greater number of Bank projects and designed to allow rigorous testing of the impact of electrification.

[] While stimulation of productive enterprise is claimed to be among the benefits of electrification, few studies have tried to quantify these

benefits using an impact evaluation methodology.


17/156

17

The methodology can be used beyond the specific PRODUSE exercise and adapted to evaluate the impacts ofany other electrification project, at relatively modest costs.

Thus the cost can be covered by most electrification projects from their M&E budget. This is essential to allow

for a wider use of such impact evaluation modules in electrification projects (which should be an explicit

element of SE4All roll-out), with the aim to produce a broader body of evidence-based findings on the energy-

poverty nexus. Depending on the available M&E budget of the project, co-funding from evaluation depart-

ments or research networks can be complementarily considered to meet the financial requirements to

implement a robust impact study.

The main objective of the PRODUSE study is (a) to gain insights into the interaction between electricity access

and productive electricity usage and (b) to improve the available toolkit for evaluation of electrification projects

(with a particular focus on productive usage) . Based on a better understanding of how modern energy access

might lead to economic and social development, effective interventions can be planned to complement en-

ergy programmes.

In parallel to this study, GIZ and the EUEI Partnership Dialogue Facility have reviewed the experience with

ongoing and past productive use promotion efforts. Based on this review, a practical guide for project plan-

ners was developed to design concrete activities to promote sustainable productive use of electricity. This

parallel guidance note has been published under the title Productive Use of Energy (PRODUSE) A Manual for

Electrification Practitioners.9

The main part of this report presents the application of the developed tools and particularly the results

regarding objective (a). Chapter 2starts with an overview of the existing literature showing that, while a

couple of solid publications exist on electrification and its impacts, rigorous project evaluations focusing on

electricity take up in general and productive use impacts are rare. Chapter 3outlines the methodology toevaluate the impacts of productive electricity usage, which is then applied in Chapters 4-6 for the case of

Benin, Ghana and Uganda. Chapter 7contains concluding remarks.

Regarding objective (b), a hands-on step-by-step guide for electrification practitioners on how to implement an

impact-based M&E systemis presented in the Annex. This tool helps project managers to move from simple

monitoring and reporting of numerical targets to a methodologically sound evaluation of the impact achieved

through electrification among micro-enterprises.

9) See EUEI PDF/GIZ 2011. The PRODUSE Manual is available at www.produse.org/manual.


18/156

18


19/156

19

Chapter 2:The Impact of Electricity Access

on Economic Development:

A Literature Review

By Benjamin Attigah and Lucius Mayer-Tasch 10


20/156

20

1. Introduction

Policy makers around the world believe that access to modern energy (both electrical and non-electrical) is anecessary requirement for sustainable development. This belief is based on three basic arguments, which

often appear in non-empirical literature on energy for development:

(i) modern energy may be a crucial input to achieving several of the Millennium Development Goals

(MDGs)

(ii) modern energy use may enable the poor in developing countries to engage in improved or new income-

generating activities (often called productive use of energy, as opposed to consumptive use), thereby

eventually leading to an improvement in their living conditions (Practical Action 2012, UNDP/WHO

2009, DFID 2002, UN 2002, UN Millennium Project 2005, Brew-Hammond and Kemausuor 2009) and

(iii) exclusion from modern energy might be a direct indicator of poverty based on definitions which refer

to living standards for instance, access to electricity is included in a recently published Multidimen-

sional Poverty Index by the UNDP (2010).

Of all modern energy types, electricity access is included most frequently as an explicit objective of national

development strategies. Hence, the focus in this chapter is on access to electricity.

Empirical evidence which can be used to validate the arguments above is surprisingly scarce. In particular,

little direct evidence has been published to underpin the second argument, i.e. the claim that electrification

can reduce poverty through enabling productive uses of electricity (IEG 2008, Kooijman-van Dijk 2008, ADB

2005, Meadows et al. 2003, Martinot et al. 2002). Moreover, quantitative evidence of the impact of electricity

on economic development (especially in comparison to other publicly provided services) hardly exists. Strong-

er evidence is needed for better-informed policy decisions, such as the priorities of public investment options

(World Bank 2010).

The few studies that do exist on the topic often lack a reliable methodology (Meadows and Riley 2003). ADB(2005) and Estache (2010) present two recent reviews of academic literature on the impact of infrastructure

on poverty reduction: both conclude that most existing studies on electrification impacts are of uncertain

value due to a series of shortcomings in the applied methodologies, such as a lack of control groups and/or

before-after data and a general failure to track the effects on poverty over a long enough time period. 11

The lack of robust evidence to date can partly be attributed to the fact that electricity is a quintessential in-

termediate good. Electricity does not represent an end in itself: it is an input factor to a large set of activities

(uses) that can improve welfare, increase productivity or generate income. The complex interactions and

synergies between multiple development factors, including other infrastructure investments next to electric-

ity and enabling political, socio-economic and cultural conditions, pose major methodological challenges to

isolating and quantifying the impact of electrification. Indeed, it is increasingly recognised that certain com-

plementary inputs or services such as business development services (BDS) or access to finance canincrease the chances that access to electricity leads to significant income generation and poverty alleviation

(ADB 2005, IEG 2008, Motta and Reiche 2001, Peters et al. 2009). However, knowledge about the extent to

which these complementary factors contribute to improving the impacts of energy investments on poverty

reduction and under which circumstances is at best incomplete (Kooijman-van Dijk 2008).

The debate on the precise role of electricity in economic development, thus, remains disputed.

This chapter provides a short review of relevant recent literature in order to better understand the contribu-

tion of energy (in particular, electricity) to economic growth and development. It looks at both macro and

micro-level research analysing the links between energy and development. The chapter will discuss only

in-depth qualitative and quantitative research. The fairly large number of policy papers citing purely anecdo-

10) The authors thank Anna Brderle, Mike Enskat, Elizabeth Elizondo, Nadja Kabierski-Chakrabarti , Sophia Kamarudeen, Nicolas Korves, Jrg

Peters, Kilian Reiche and Peggy Schulz for valuable inputs.

11) Some recent studies have begun to apply more comprehensive research methods (e.g. World Bank 2009a, World Bank 2009b). Nevertheless,

these studies are still few in number and they represent mostly grey literature which is not yet established in more academic research.


21/156

21

tal evidence on productive use of electricity is not taken into account here. The review focuses on the role ofelectricity access; questions of service quality or reliability of energy supply will only be considered as far as

they influence the uptake and impact of (newly provided) energy access. In line with the general focus of the

PRODUSE Study, the chapter takes a one-dimensional perspective on development with economic parame-

ters like income generation, growth and productivity, inter alia, as the best measurable development indica-

tors. Impacts on education or health for example obviously form important aspects of development but

their measurement requires more complex methodologies and therefore they are not discussed here.

The remainder of this chapter is structured as follows: the next section provides a basic conceptual back-

ground by outlining the steps that lead from energy supply to poverty reduction. Section 3 discusses some

methodological issues. This is followed by an overview of the existing empirical evidence at the macro-level

(Section 4) and at micro-level (Section 5). Section 6contains concluding remarks.

2. Conceptual Background:The Steps from Electricity Supply to Poverty Reduction

As Kooijman-van Dijk (2008) points out, one of the reasons why there is little understanding of the links

between electricity supply and poverty reduction through income generation is because the relationship con-

sists of several steps and many factors influence each of these steps. The first step towards a business benefit-

ing from electricity supply is the physical provision of electricity and the entrepreneurs decision to make use of

it. However, it is the steps that follow, namely the actual use of electricity and the subsequent changes that

electricity use brings in the enterprise (e.g. increased productivity), which can ultimately lead to impacts at

enterprise level, such as increased income. The theory regarding the causal chain from energy supply infrastruc-

ture to development outcomes is displayed in Figure 1(adapted from Kooijman-van Dijk 2008: 6). This concept isdiscussed in more detail in the methodology developed for the empirical research of the study at hand.

Obviously this figure shows only a hypothesisof interaction and direction of impact. However, it provides a

useful framework for further analysis, as the structure makes clear the relationships between the different

variablesthat are typically analysed in the literature (energy supply quality and reliability of energy supply

- energy consumption/use productivity growth/GDP/income (inequality) poverty reduction). For pur-

poses of empirical investigation, these variables can in turn be measured by different indicators. The empirical

findings presented in subchapters 4 and 5, for both macro and micro-level, are structured along these lines.

3. Methodological IssuesEvidence of the contribution of energy to economic development is often presented in the form of simple cor-

relationsbetween electricity and welfare indicators such as GDP or the Human Development Index (HDI) at the

macro-level (e.g. IEA 2004) or household income at the micro-level. Such correlations are then presented as evi-

dence that energy causes positive development outcomes. For example in a study by White (2002: 34) the figure

below is presented as evidence that human development responds dramatically to initial electricity additions.

It is important to point out, however, the simple but all too often neglected fact that correlation does not

imply causality! In our example, it is just as plausible that improvements in the HDI lead to increases in energy

consumption and not just the other way around.

Academic research commonly uses regression analysisto test the magnitude and direction of causal relation-

ships between variables in a data set.12In the literature on the link between electricity (or on a more generallevel, infrastructure) and economic development, the main explanatory variable is usually either electricity

consumption or electricity supply. The dependent variables analysed in the literature are productivity, output,

growth, income (inequality), employment or poverty reduction.


22/156

22

Figure 1: The Theory: Steps from Electricity Supply to Poverty Reduction

ECONOMICDEVELOPMENT

ENERGY(ELECTRICITY

)

Poverty reduction

Source: adapted from Kooijman-van Dijk (2008)

Source: White (2002)

Energy supply (adequate qualityand quantity)

Generation of income or Gross

Domestic Product (GDP);

other benefits (comfort of operation)

Energy consumption/use;

adoption of appliances

(i) Changes in enterprise:

Changes in productivity,

quality, cost of production,

prices, volumes of products,operating hours, employment;

creation of new products or

services

(ii) Creation of new enterprises

Figure 2: Macro-Level Correlation Between Electricity and Human Development

0

0 2000

Electricity per capita (kWh/capita)

4000 6000 8000

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1

HDI

Low Human

Development

Medium

Human

Development

High Human

Development

Increaseinenerg

ydemand


23/156

23

A common methodological framework to research the electricity-development link is the production func-

tion. Within the framework of the production function, the impact of electricity on (aggregate) output is usu-

ally modelled in two main ways: firstly, directly when electricity services enter production as an additional

input and secondly, indirectly when they raise total factor productivity by reducing transaction and othercosts, thus allowing a more efficient use of conventional productive inputs (Straub 2008a). Examples for such

indirect effects of electricity infrastructure are: (i) well-maintained infrastructure may reduce operating costs

of private capital or extend its life span, such as machines connected to stable voltage lines, (ii) high-quality

infrastructure can reduce private adjustment costs to unreliable electricity services, e.g. investments in power

back-up systems, (iii) it can increase labour productivity, e.g. through a more efficient structuring of business

processes as a result of improved information and communication technology. It is important to point out,

that adverse effects of investments in electricity infrastructure may also occur, e.g. when public investment

crowds out private investment.

4. Macro-Level Research

4.1. Introduction

The macro-level literature analysing the link between electricity and economic development so far remains

limited, as the mainstream economic literature on growth and development pays little attention to the role

of energy (Estache 2010). There are, however, a number of studies from the field of energy economics that look

at the causal relationship between energy consumption and economic growth. A second distinct body of

literature, as mentioned in the section above on methodological issues, focuses on investigating the impacts

of infrastructure investments on a variety of development outcomes (such as growth, productivity or poverty

reduction). Most of these studies include energy infrastructure as one variable of interest.13

In the context of the aforementioned body of literature, it will be of interest to not only look at the direct im-pacts of energy infrastructure on different development outcomes, but also to analyse its impact compared

with other publicly provided infrastructure services. Such an assessment will be relevant from a policy per-

spective as it allows policy makers to better evaluate and prioritise different types of investments in order to

allocate public funds as efficiently as possible.

4.2. Empirical Evidence

This section surveys the empirical evidence on the impact of electricity on different development outcomes,

manifested in three broad categories of impacts: productivity, growth and poverty.

Impacts on Productivity

The literature examining the development impact of infrastructure, most of which includes electricity infra-

structure as one variable of interest, started with the seminal work of Aschauer (1989). He finds that the stock

of public infrastructure capital including electricity is a significant determinant of aggregate total factor

productivity (TFP). His results suggest that infrastructure played an important role in the productivity slow-

down in the U.S. which started around 1973. Earlier studies exploring this phenomenon had ignored the role

of infrastructure and focused on other factors such as energy prices or R&D (Gramlich 1994). Critics of Aschauers

12) Regression analysis is a method for numerical data analysis where the relationship among the variables in a data set is summarised as

an equation. In this equation the variable of interest, or the dependent variable, is expressed as a function of one or several explanatory

variables.

13) Infrastructure typically includes energy, transport, telecommunications, water, irrigation and sanitation.


24/156

24

work pointed out that the economic significance of his results was considered implausibly large and that he

failed to address several methodological issues.14Later studies applied more sophisticated econometric tech-

niques to correct for these methodological problems (see Gramlich 1994, Romp and de Haan 2005 and

Estache and Fay 2009 for an overview of these studies).

Table 1summarises the studies reviewed in this section, which look at electricity as a variable of interest next

to other infrastructure services like transport and telecommunications. Different electricity indicators are

employed, including electricity generation, electricity generation capacity and investment in electricity infra-

structure. The overall effect of electricity on productivity varies across countries. Positive effects of electricity

on productivity are found in various geographic areas (Fedderke and Bogetic 2006, Noumba Um, Straub and

Vellutini 2009), while only insignificant or even negative impacts also emerge for some other regions (Fan,

Zhang and Zhang 2002, Fan, Hazell and Thorat 1999, Straub, Vellutini and Warlters 2008). The evidence also

shows that in some countries such as China, India and Thailand electricity displays a smaller productivity

effect than other infrastructure investments, notably agricultural research and development.

14) The most important issue concerned the potential of reverse causation from public capital to productivity and output. Neglect of this

potential endogeneity is likely to cause an upward bias in the estimated returns to infrastructure (Romp and de Haan 2005).

Source Country/ Region Output Indicator(s) Conclusion

Edquist and

Henrekson (2006)

Germany,

Sweden, UK, US

Rate of productivity

growth

Productivity growth occurs with a distinct time lag

following electrification (about 4050 years for

electrification and the ICT revolution and about 140

years for the steam engine). No clear evidence of

high productivity growth rates for both electric

machinery industry and the steam engine

producing industry was found.

Fan, Hazell and

Thorat (1999)

India Agricultural

productivity

Additional government spending on rural

electrification has low productivity effects.

Government expenditure on rural roads andagricultural research and extension promote

greatest growth in agricultural productivity.

Fan, Jitsuchon and

Methakunnavut

(2004)

Thailand Agricultural labour

productivity

Investments in rural electrification have the second

largest impact on agricultural productivity growth

after agricultural research and development.

Fan, Zhang and

Zhang (2002)

China Agricultural

productivity

No significant effect of electricity on agricultural

productivity. Agricultural research has largest effect

on productivity.

Fedderke and

Bogetic (2006)

South Africa Labour productivity

and

TFP growth

Electricity generation is positively related to labour

productivity and TFP growth.

Noumba Um,

Straub and

Vellutini (2009)

North Africa and

Middle East

TFP Electricity production helps explain cross-country

differences in TFP growth.

Straub, Vellutini

and

Warlters (2008)

East Asia TFP growth No significant contribution of electricity generating

capacity. Indonesia (as a relatively poor country) is

the only exception and shows negative impacts.

Table 1: Effects of Infrastructure and Energy on Productivity

Source:adaptedfrom

Pinstrup-AndersenandShimokawa(2007)


25/156

25

Impacts on Growth

The literature on electricity and growth can be split into (a) the energy economics literature analysing the

causal relationship between electricity consumption and growth and (b) the literature on infrastructure anddevelopment which often includes electricity infrastructure as a variable of interest. Studies in the first cate-

gory, which analyse the relationship between electricity consumption and GDP growth, produce conflicting

results in terms of the existence and direction of causality between the two variables. This conclusion is

drawn by Ozturk (2010) who has undertaken an extensive review of this body of literature of more than a

hundred studies from a wide range of countries, including both country-specific and multi-country analyses,

covering the period 1978 to 2009 and applying a variety of methodological approaches. Ozturk distinguishes

between four types of relationships: no causality, uni-directional causality running from economic growth to

electricity consumption, uni-directional causality running from electricity to growth and bi-directional cau-

sality between economic growth and electricity consumption.

With regard to methodology, Ozturk suggests in line with Karanfil (2009) that researchers should use more

appropriate econometric techniques in the future, as the methods most often applied to date are of limitedvalue with regard to the issue in question. The traditional methods (i.e. ordinary least squares) will not yield

the required insight but rather increase the number of conflicting results and cast doubt on the reliability of

their policy recommendations. A number of recent studies have sought to apply more comprehensive econo-

metric techniques and address key methodological issues. For instance, in his study on the nexus between

electricity supply, employment and real GDP in India, Gosh (2009) makes a case for electricity supply being a

more meaningful indicator than electricity demand in countries with high levels of non-technical transmis-

sion and distribution losses (e.g. as a result of theft or pilferage of electricity), as the use of official data may

lead to a systematic underestimation of real electricity consumption. Next to findings in relation to employ-

ment effects of electricity, the author establishes short-run causality running from growth to electricity

supply (based on use of electric appliances in the industrial, commercial and domestic end-use sectors) but

finds no causality running from electricity supply to real GDP.

The second category of studies that examine the electricity-growth-nexus, i.e. those that try to quantify the

contribution of different kinds of infrastructure to income and growth, find mostly positive effects of electric-

ity on economic growth. In a recent survey of the literature on infrastructure and growth in Africa, Foster and

Briceno-Garmendia (2010), conclude that there are strong indications of a positive impact of infrastructure on

growth. Several of the reviewed studies include electricity in their estimations and show a beneficial growth

effect of electricity (e.g. Ayogu 1999, Caldern and Servn 2008, Estache, Speciale and Veredas 2005).

Table 2summarises a number of empirical studies that examine the effects of electricity infrastructure on growth.

As can be seen from the table, only a few studies report relatively small impacts of electricity compared to

other infrastructure investments (Fan, Zhang and Zhang 2002) or fail to find significant impact of electricity

on growth (Straub, Vellutini and Warlters 2008). Despite the evidence pointing towards positive growth

effects of electricity infrastructure, some authors suggest to interpret such results with caution. For instance,Straub, Vellutini and Warlters (2008) state that results from aggregate macro-level data should generally be

interpreted with care. They point out that the primary function of infrastructure investments may not be to

directly promote economic growth but rather relieve constraints and bottlenecks to growth as they arise.

Other authors such as Ayogu (2007) voice more fundamental concerns about the infrastructure-growth

literature. Ayogu conducts an extensive review of the empirical literature and he concludes that the question

of whether infrastructure matters for growth has not been satisfactorily resolved. Moreover, in his view this is

only a marginally important issue, way beyond what could be the value added from totally resolving the

issue. Instead, the author proposes to focus future research efforts on much more relevant policy issues, such

as analysing how much infrastructure matters exactly in different contexts.


26/156

26

Source Country/ Region Output Indicator(s) Conclusions

Ayogu (1999) Nigeria GDP Strong association between infrastructure,

including electricity and output in panel data

was found.

Binswanger,

Khandker and

Rosenzweig

(1993)

India Agricultural invest-

ment and crop output

Electrification has a clear effect on agricultural

investment (e.g. in pumps) and thereby also on

agricultural output.

Caldern and

Servn (2008)

Africa GDP growth per

capita, Gini coefficient

Infrastructure stocks as well as infrastructure

service quality have positive impacts on long-run

growth and income equality - electricity is included

in the estimations as part of aggregate infrastruc-

ture indices.

Caldern (2009) Africa Growth in GDP per

capita

With Mauritius infrastructure Africas growth per

capita would be enhanced by 2.2 percent per year.

African countries would gain more from larger

stocks of infrastructure than better quality of

existing infrastructure. The largest payoffs are for

telephone density, electricity-generating capacity,

road network length and road quality.

Canning and

Pedroni (2004)

Various GDP per capita Long run effects of investment in electricity

generating capacity are positive in a large number

of countries, with negative effects being found in

only a few.

Easterly and

Levine (1997)

Various GDP growth rate Infrastructure, measured by telephones per worker

and electricity losses, is strongly and significantly

correlated with growth. However, no significant

impact is found when measured as electricity

generating capacity.

Esfahani and

Ramirez 2003

Various GDP growth per

capita

Substantial impact of infrastructure, measured by

electricity and telecommunications infrastructure,

on GDP growth; this impact in turn depends on

institutional and organisational capabilities.

Estache, Speciale

and

Veredas (2005)15

Africa GDP per capita Roads, power and telecommunications infrastruc-

ture but not sanitation contribute significantly

to long-run growth in Africa.

Fan, Zhang and

Zhang (2002)

China Agricultural growth Electricity has a positive effect on agricultural GDP

but much weaker one than the other factors

analysed

Noumba Um,

Straub and Vellu-

tini (2009)

Middle East and

North Africa

GDP growth per

capita

Impact of growth of electricity production on GDP

growth. MENA countries demonstrate lower

returns than developing countries as a whole,

probably due to higher levels of investment and

subsequent diminishing returns.

Seethepalli,

Bramati and

Veredas (2007)

East Asia GDP per capita Positive and significant effects for electricity

on growth were determined.

Straub, Vellutini

and

Warters (2008)

East Asia GDP growth per

capita

No robust impact of electricity production on

growth was detected.

Table 2: Effects of Infrastructure and Electricity on Economic Growth

Source:adaptedfrom

FosterandBriceno-Garmendia(2010)andPinstrup-AndersenandShimoka

wa(2007)


27/156

27

Source:adaptedfrom

Pinstrup-AndersenandShimokawa(2007)

15) It should be noted that the reliability of this finding is questionable because the control group of non-electrified households was very

small (31 households, compared to 1,012 electrified households) and the authors

Impacts on Poverty

There are several empirical studies that focus on the question of whether increased electricity access actually

benefits the poor or whether it tends to increase incomes of the upper income strata disproportionately. Inthis context it will be of interest to specifically compare the impact of energy infrastructure with the impact

of other infrastructure services. This might allow policy makers to evaluate and prioritise infrastructure

investments in order to allocate public funds as efficiently as possible.

A general observation from these studies is that electricity has a relatively small effect on poverty as com-

pared to other infrastructure investments, notably roads. Table 3provides an overview of the cited studies on

the nexus between electricity infrastructure and poverty reduction.

Fan, Hazell and Thorat (1999), for example, using data for 1970 to 1993 from India, conclude that government

spending should focus on rural roads and agricultural research and extension, as these types of investments

have the greatest poverty impact (i.e. the number of people raised above the poverty line for each additionalmillion rupees spent). Regarding rural electrification (as well as irrigation), they state that additional govern-

ment spending has no discernible impact on poverty reduction.

One exception to these findings is a study by Fan, Jitsuchon and Methakunnavut (2004) on Thailand. Their

results show that out of different types of public investments (agricultural R&D, irrigation, rural education,

road infrastructure and electricity infrastructure), investments in rural electrification have the largest poverty

reduction impacts. The authors suggest that this differing result is due to Thailands status as a middle-

income country. They state that in lower-income countries returns from road investments usually are higher

than from electricity or telecommunications. However, as Thailand had already invested heavily in rural roads,

additional investments in roads will only yield diminishing returns. This can explain why in the case of Thai-

land the returns on investment in electricity are higher than for investments in roads.

Table 3: Effects of Infrastructure on Poverty Reduction in Developing Countries

Source Country/ Region Output Indicator(s) Conclusions

Balisacan 2001 Philippines Proportion of the ruralpopulation living

below the provincial

poverty line

No significant effect of electricity access;roads have highest impact.

Fan, Hazell and Thorat

(1999)

India Number of poor

reduced per million

rupees infrastructure

investment

Additional government spending on

rural electrification has no discernible

impact on poverty reduction. Spending

on roads has largest impact on poverty

reduction, followed by agricultural

research.

Fan, Jitsuchon, and

Methakunnavut (2004)

Thailand Number of poor

reduced per million

bahts infrastructure

investment

Among different public investments

(agricultural R&D, irrigation, rural

education, road infrastructure) invest-

ments in rural electrification have the

largest poverty reduction impacts.

Fan, Zhang and Zhang

(2002)

China Number of poor

reduced per 10,000

yuans infrastructure

investment

Positive effects of infrastructure

investments in rural electrification,

which are however smaller than those

of investments in rural education,

agricultural research and roads.


28/156

28

5. Micro-Level Research

5.1. Introduction

As far as impacts of electrification on the micro-level are concerned, the empirical research has taken different

methodological approaches, looking at different units of analysis. There are a number of energy-specific stud-

ies and general enterprise surveyslooking at various types of businesses (formal and informal, small and large)

and household surveysanalysing economic indicators such as income from home businesses among other

impacts of electrification.

Impacts of electricity on the micro-level are often examined using the same indicators as on the macro-level

(enterprise creation, business activity, firm productivity, employment, income (equality), gender and poverty

reduction) and the conceptual framework of the second subchapter also applies here. The main difference lies

in the level of aggregation.

Besides a key methodological weakness of macro-level research, very few micro-level studies so far go beyond

showing correlations by attempting to employ rigorous methods that are suitable for proving electrification

impacts on MSMEs by providing robust evidence for a causality between electrification and MSME performance.

This subchapter will first discuss which factors have been found to influence whether and how electricity

impacts on enterprise performance emerge. It will then review the literature that provides evidence on

impacts of electricity on enterprises, looking at different forms in which such impacts can be measured on the

firm and household level: the creation of new businesses, productivity, employment, poverty and income of

businesses and households.

5.2. Factors Influencing Impacts of Energy

The uptake of electricity (i.e. the decision for connection and the magnitude of kWh use) and the impacts of

electricity use on MSMEs depend on various external and internal factors including access to markets (inter-

national, national and local), company location, income levels in the local economy, quality of supply and

financial as well as other assets of the entrepreneur/firm.

Using an adapted version of the Sustainable Livelihoods Framework, Kooijman-van Dijk (2008) distinguishes

between financial, physical, human and social assets that influence an entrepreneurs ability and willingness

to connect to and use electricity. The same assets influence if/how electricity is used for productive activities

which translate an electricity connection into economic benefits. She also differentiates between the strate-

gies pursued by entrepreneurs who were forced to engage in a certain (non-farm) income-generating activity

due to a lack of other opportunities (coping strategies) and those with a clear growth orientation (accumula-

tion strategies).

Quality and reliability of electricity supply is an important factor both for the decision to connect and for the

impact on MSME performance. In some countries the reliability is so low that electricity-reliant businesses

have no choice but to invest in diesel generators if they want to maintain business operations at a minimum

level of steadiness. Foster and Steinbuks (2009) estimate that generators owned by firms account for about

6 % of total installed generation capacity in Sub-Saharan Africa and up to 20 % in low-income countries.

According to the World Banks Doing Business report, firms in low-income countries are affected by electricity

supply interruptions on average 18 times in a typical month. The resulting workflow interruptions and the

damage of sensitive electrical equipment such as computers caused by voltage fluctuations can curtail prof-

its significantly. Business managers interviewed for the Doing Business project in the various countries esti-

mated that losses due to electricity outages amount to an average of 3.2 % of annual sales and as much as

22.6 % in Malawi (World Bank 2010). Studies on the impact of unreliable power supply on firm productivity are

reviewed in the subchapter Impact on productivity below.


29/156

29

5.3. Empirical Evidence

Impact on Creation of Enterprises

The creation of new, often informal (home) businesses triggered through access to electricity has been

analysed in a number of countries using data from household surveys. Some of these studies do find positive

correlations between electrification and (increase in) numbers of MSMEs; however, such results must be

interpreted with care, as prioritisation of economically dynamic areas for electrification can easily result in a

bias among the surveyed treatment (i.e. electrified) and control (non-electrified) areas. For example, an ESMAP

study conducted in the Philippines found that across four provinces, 25 % of the households in electrified

areas are running a home business (mainly small retail shops but also tailoring etc.) compared to 15 % in

non-electrified areas and that the variety of these businesses is greater in electrified areas (ESMAP 2002).

However, as pointed out by Kooijman-van Dijk (2008), it does not show whether this was a result of electrifica-

tion or whether the electrified areas where selected precisely because of more favourable socio-economic

characteristics in the target area.

The reviewed studies on business creation through electricity access are summarised in Table 4.

Table 4: Effects of Electricity Access on Number of Businesses

Source Country/Region Data Source/Sample Size Conclusion

Arnold, Mattoo

and Narcisco

(2008)

10 African

countries

Approx. 1,000

manufacturing

enterprises

Unreliable electricity supply has a significant

negative impact on a firms total factor

productivity, while generator possession has

a significant positive effect.

Barnes and

Binswanger

(1986)

India Surveys conducted in 108

villages in 1966 and 1980

Rural electrification had a direct impact on

agricultural productivity through private

investment in electric pumps.

Blalock and

Veloso (2007)

Indonesia 20,000 manufacturing

enterprises

Significant positive effect of energy

consumption on firm productivity was found.

Eifert et al.

(2008)

17 African

countries

Enterprise surveys Indirect costs (of which energy costs comprise

the largest share) are a major factor for

explaining the low productivity of enterprises

in Africa.

Escribano et al.

(2009)

26 African

countriesInvestment climate

surveys

Infrastructure quality has

produse study_full text

Documents